Editing peptide presentation to T cells

See allHide authors and affiliations

Science  24 Nov 2017:
Vol. 358, Issue 6366, pp. 992-993
DOI: 10.1126/science.aaq1398

In mammals, two types of major histocompatibility complex (MHC) molecules, MHC I and MHC II, are found on the surface of cells in association with peptides. If these peptides are antigenic—that is, if they induce an immune response—the resulting MHC I–peptide complexes are recognized by CD8+ T cells, which are effector immune cells, whereas MHC II-peptide complexes are recognized by CD4+ T cells, which facilitate both humoral and cellular immune responses. MHC I and MHC II clasp the peptides in a binding groove consisting of two antiparallel α helices that overlay an eight-stranded β sheet. On pages 1064 and 1060 of this issue, Jiang et al. (1) and Thomas and Tampé (2) reveal how peptide binding to MHC I molecules is regulated.

Both MHC I and MHC II are subjected to intracellular “editing” that ensures that the associated peptides are bound with high affinity. An iterative process facilitated by specific protein editors results in the dissociation of low-affinity peptides and their replacement by higher-affinity peptides. For MHC II, the editor involved is human leukocyte antigen DM (HLA-DM, in humans) or H2-DM (in mice) (3). For MHC I, the exchange occurs in the endoplasmic reticulum (ER), and until recently, tapasin was the only known peptide editor. Tapasin forms part of a complex called the peptide-loading complex (PLC), which includes the MHC I molecule undergoing peptide exchange and the transporter associated with antigen processing (TAP), which delivers peptides generated by the proteasome into the ER from the cytosol. Also included in the PLC are two “housekeeping” proteins, calreticulin and ER protein 57 (ERp57), which facilitate the folding of newly synthesized glycoproteins in the ER (3).

Recently, the understanding of MHC I peptide editing was complicated by the identification of TAP-binding protein–related (TAPBPR), a homolog of tapasin that shares its peptide editing function (4). TAPBPR does not associate with the PLC, and it remains unclear where it operates within the cell. However, in vitro analysis of its capacity for peptide editing proved to be much simpler than for tapasin, which must either be purified as a disulfide-linked dimer with ERp57 or artificially coupled to MHC I by means of a leucine zipper domain (5, 6). Peptide exchange by tapasin-ERp57 heterodimers requires calreticulin, which simultaneously interacts with the N-linked glycan moiety (a sugar adduct) of MHC I and with ERp57 (3). Further complicating the issue, the N-linked glycan must terminate in a glucose residue, which signifies a transient state that is also required for the calnexin-calreticulin folding cycle. Peptide exchange by TAPBPR requires neither ERp57 nor calreticulin, and a version of TAPBPR that lacks the transmembrane region can mediate peptide exchange by soluble MHC I molecules (7, 8).

The mechanism of MHC I peptide editing

TAPBPR-induced structural changes in MHC I select for high-affinity peptide binding. High-affinity peptides stabilize the MHC I–peptide complex such that the interaction with TAPBPR no longer occurs. This results in the accumulation of a population of MHC I molecules containing high-affinity peptides, a process termed peptide editing.


The structural explanation for HLA-DM–mediated peptide exchange by MHC II molecules was uncovered in 2012 (9). HLA-DM was cocrystallized with a soluble MHC II molecule, HLA-DR1, containing an N-terminally truncated peptide that was disulfide-linked to a cysteine residue introduced into the MHC II binding groove. They found that major conformational changes occurred within the binding groove. A similar explanation for the function of TAPBPR and MHC I peptide exchange now exists. Jiang et al. (1) used an approach like that used to generate MHC II–HLA-DM complexes (9). MHC I molecules normally bind an 8– to 10–amino acid peptide and are associated with a nonpolymorphic protein, β2-microglobulin (β2m). Rather than using peptide-free MHC I–β2m dimers, which are unstable, they searched for an MHC I molecule amenable to the introduction of C-terminally truncated antigenic peptides that could be disulfide-linked to a cysteine residue introduced into the floor of its peptide-binding groove. They demonstrated that the mouse MHC I molecule H2-Dd could bind recombinant human TAPBPR. One such complex was purified and crystallized, and a structure was obtained. Thomas and Tampé (2) used an alternative strategy that relied on producing soluble MHC I molecules (H2-Db) containing a noncovalently bound photocleavable peptide that allows the transient formation of peptide-free MHC I by irradiation with ultraviolet light (10). By irradiating H2-Db–β2m dimers containing such a peptide in the presence of recombinant human TAPBPR, they were able to generate H2-Db-TAPBPR complexes that were amenable to crystallization and structural analysis. In each case, the complexes obtained were trimeric, consisting of a mouse MHC I protein, human β2m, and human TAPBPR.

Although there are some structural differences, both groups conclude that the critical effect of the TAPBPR interaction is to widen the peptide-binding groove in MHC I. A segment of one of the α helices in MHC I, called α2-1, is moved out and down from its normal position as a consequence of a direct contact with TAPBPR. There is also a downward shift of one of the β strands that form the floor of the groove. Thus, TAPBPR stabilizes a conformation that cannot associate with a conventionally bound peptide. The membrane-proximal MHC I α3 domain and β2m also interact with TAPBPR but are shifted compared to their positions in peptide-bound MHC I molecules. TAPBPR itself has a very similar structure to tapasin, as expected from their homology, and the region of its interaction with MHC I is as predicted by mutational analysis of tapasin and TAPBPR (8, 11). The slight differences between the H2-Dd- and H2-Db-TAPBPR structures may be because they are the products of different genes or because of the retention of the covalently associated pentameric peptide in the former and the likely absence of any peptide in the latter. In H2-Dd-TAPBPR, the attached short peptide is not visible, suggesting that it is mobile as a consequence of the TAPBPR interaction. In H2-Db-TAPBPR, a TAPBPR loop containing a short α helix, which is not visible in the H2-Dd-TAPBPR structure, interacts with the α2-1 helix of the peptide-binding groove where the C-terminus of an associated peptide normally binds. Thomas and Tampé refer to this as the “scoop loop,” and the loop interaction is incompatible with the presence of a peptide. Overall, both structures point to a mechanism of peptide exchange that relies on the inability of TAPBPR and a peptide to simultaneously associate with MHC I. TAPBPR interaction with MHC I that contains a low-affinity peptide causes a conformational change in the binding groove that releases the peptide; this interaction will not occur if a high-affinity peptide is bound (see the figure).

Therefore, tapasin and TAPBPR appear to promote peptide dissociation from MHC I in the same way. Tapasin is part of the PLC, whereas the functional context for TAPBPR is unknown. Recently, TAPBPR was found to physically associate with the enzyme UDP-glucose:glycoprotein glucosyltransferase 1 (UGT1) (12), which is required to maintain monoglucosylation of the N-linked glycan of MHC I that is required for its incorporation into the PLC (13). TAPBPR may target UGT1 to the MHC I glycan after peptide dissociation. The addition of glucose to the N-linked glycan could induce the association of the peptide-free MHC I molecule with the PLC, where binding and editing of peptides specifically transported from the cytosol may occur. Further work should investigate if this is the case or whether the function of TAPBPR is independent of the more conventional tapasin-mediated pathway of MHC I peptide loading.


View Abstract

Stay Connected to Science

Navigate This Article